Fixing Bugs Systematically
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and acts upon data from external sources within the workspace.
- Ingestion points: The agent is instructed to read various project files (e.g.,
docs/feature-spec/,docs/user-stories/,docs/api-contracts.yaml) and await responses from delegated agents in./agent-responses/. - Boundary markers: The instructions do not define any delimiters or provide guidance to ignore instructions embedded within the processed documents or agent outputs.
- Capability inventory: The skill utilizes file reading (
read_file,grep), file modification (implementation of fixes and documentation updates), and the ability to spawn and manage auxiliary agents. - Sanitization: There is no evidence of sanitization, filtering, or validation of the content retrieved from these external ingestion points before it is incorporated into the agent's context and logic.
Audit Metadata